PROJECT SUMMARY/ABSTRACT Copy number variation (CNV), too many or too few copies of a segment of the genome, underlies many important human medical issues. We had predicted that, based on our MMBIR model for the generation of much CNV by aberrant repair of broken replication forks, that there would be a unidirectional tract of hypermutation of considerable length extending from the junction of the CNV. We devised a series of techniques to analyze seven million base-pair tracts of DNA sequence surrounding CNVs on chromosome 17. With these techniques we were successful in determining the precise structure of, and mutations linked to, 26 CNVs. We also analyzed both parents' genomes. We discovered what we sought: there are tracts of hypermutation in one direction linked to the CNV extending for up to one million base-pairs from the CNV. This confirms that our postulated mechanism is responsible for at least half of these CNV events. For the other half that do not show hypermutation, we will obtain data from a larger sample of parents until we can say whether these arose by a different mechanism or whether it is the tail of the distribution of the same mechanism. Because of the very highly detailed resolution of our analyses, we are able to see into the mechanisms that generate the hypermutation tracts. We find two mechanisms that we can provisionally decipher and possibly a third whose cause we have not yet found. Near the CNV, a low processivity polymerase makes multiple template switches and slips on the template that is being replicated. Further away, we see evidence of processes acting on single-stranded DNA giving clustered mutations of a unique signature. The third signature might relate to the diminished mismatch repair that is expected when broken replication forks prime replication. The additional data will make the signatures clearer and allow us to determine the causes. The next step is to generalize the findings to the rest of the genome. We will do this by whole genome sequencing of CNVs at other sites. In a third project we are going to use new sequencing technology to decipher the recurrent CNVs that arise by crossing-over between repeated sequences. This has been an intractable problem with previous technology because of the numerous copies of the sequence present in the cell. We expect to find the precise positions of crossovers and gene conversion tracts, indicating the rules that govern where they will fall, and determine the extent and signature (and therefore the cause) of any new mutations. Together these Aims will extend understanding of DNA repair events gone wrong that lead to genomic disorders, and potentially suggest ways to control or avoid this happening.